Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 111 (2017) 885 – 894

8th International Conference on Sustainability in Energy and Buildings, SEB-16, 11-13 September
2016, Turin, ITALY

A numerical case study: Bovet approach to design a Francis turbine

runner
Eyup Kocaka, Salih Karaaslana, Nuri Yucela,* , Furkan Arundasa
a
Gazi University, Faculty of Engineering, Mechanical Engineering Department, Ankara, 06570, Turkey

Abstract

In this study, analytical calculation and numerical simulations were performed to design a Francis turbine runner blade. Single
blade was designed by using Bovet approach. Blade geometry was created by using Ansys Bladegen V16.1. Grid model has been
generated by using Ansys Meshing, with using fine mesh quality, tetrahedron type mesh. k- Ɛ turbulence model have been selected
for steady state analysis. Pressure distributions are obtained on meridional plane and also on suction and pressure sides of the blade.
The results of the numerical analysis showed that, Bovet design approach is able to calculate a runner that has an efficiency of 1%
different with respect to committed efficiency. At the inlet of the suction side, the negative pressure is observed, that can cause
cavitation and oscillation.

© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of KES International.

Keywords: Type your keywords here, separated by semicolons ;

1.
Introduction

In recent years; due to the decline of fossil fuels, there is a tendency to renewable energy sources. One of the biggest
and also well-known recipe of producing green energy is hydroelectric power plants. According to U.S. Geological
Survey, the percentage of producing energy from hydro turbines is %16 in the world [1]. There are generally five types
of hydraulic turbines are used in Hydraulic Power Plant technology. Hydraulic turbines are classified due to head and
flow rates. Figure 1 shows that the selection of hydraulic turbines by using head and flow rates [2]. Francis turbine is
the most used hydro turbine type and it is suitable for working with wide-spectrum of head and flow rates. The most
important part of a Francis turbine is the runner that transforms the hydro energy to mechanical energy. In general

* Corresponding author. Tel.: +90-312-5823422; fax: +90-312-2319810.

E-mail address: nuyucel@gazi.edu.tr

1876-6102 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of KES International.
doi:10.1016/j.egypro.2017.03.251
886 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894

turbine runner is composed of blades which have complex geometries. Thereby, in order to get efficient turbine, design
of the runner has high priority. The blades, constructed the runner, have very complex shapes, so to design blades
analytically causes many complications.

Fig. 1. Radial turbine selection chart by net head - flow [2].

There were many studies about Francis turbine design in the literature. One of the most important design approach
of Francis turbine runner is Bovet method which uses empirical equations to obtain parameters of Francis type hydro
turbine runner [3]. In Bovet method, the dimensionless specific speed value is the main parameter to determine whole
turbine dimensions. Miloş and Bârglâzan [4] investigated the refurbishment of the hydraulic turbine’s runner, medium
head Francis by CAD technique. The authors were used Bovet approach to calculate meridional channel and blade
splines. It is seen that the authors preferred parabola arcs method to obtain leading and trailing edges profile.
Conformal mapping method is used to design the turbine runner. By rectification of the flow velocities, the flow
streamlines at the inlet of the runner is matched with the streamlines of wicket gates. Therefore, swirl at the outlet of
the runner and pressure oscillations in the draft tube are reduced. Choi et al [5] studied rehabilitation of a 500 kW
Francis turbine runner by changing internal parameters of the runner while using the old turbine components such as
spiral case, guide vanes and spiral casing. Part by part optimization method is used. It was achieved a model which
has improved cavitation characteristic with extremely smooth performance over a much wider range of operations
compared to the existing design. The improved runner and different guide vane operating conditions were tested
experimentally. According to experimental tests, the total efficiency of the turbine was increased about 9.93 percent.
Numerical simulation of cavitating turbulent flow in a high head Francis turbine at part load operation was conducted
by using OpenFOAM code in the study of Zhang and Zhang [6]. k-ω SST turbulence model was used for calculating
Reynolds averaged Navier-Stokes equations. A repetitive solver was performed while residual values of convergence
criteria were set to 10-5 grade. As a result of this study, the authors indicated that numerical approach is a suitable tool
to analyze cavitation behavior. Numerical setup which is constructed by using OpenFOAM codes was able to simulate
cavitation characteristics. Anup et al [7] conducted a study on temporary numerical validation in a Francis turbine by
using ANSYS CFX Solver. In their study k-ε with Kato – Launder correction and SST turbulence models were used
and also Large Eddy Simulation (LES) and Reynold Stress Models have been adopted for numerical approach. During
the design of Francis turbine, the important phenomenon, that is pressure fluctuation due to rotor-stator interaction
and occurrence of vortex rope in draft tube were investigated numerically at partial load operation condition [7]. Thapa
 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894 887

et al [8] searched on optimizing sediment erosion on a Francis runner blade by using CFD. The authors were used a
commercial software which is named “Khoj” to design new types of profiles. To sum up the study in a brief way;
changing beta angles of the blade will be also caused a change on erosion level on the blade. Nonetheless, it is seen
that chancing of the angle on a blade has no effect on runner dimensions. In another study, high specific speed Francis
turbine was optimized with using simplex algorithm by Obrovsky et al [9]. They were developed an algorithm which
uses simplex method to optimize blade geometry. After optimization processes were completed, they generated 3D
fluid volume of the blade via BladeGen. V14. CFD analysis of the runner and whole turbine for high specific speed
turbine were conducted for prediction of the cavitation. The optimum operational point and the maximum power output
were chosen as the points for cavitation analysis. They managed to create final design that has %100 peak efficiency
at full load flow rate by computing over 1000 variants nearly 250 computational hours [9]. Another study was carried
out by Prasad [10]. CFD validations were performed by using the experimental results of turbine. In this study,
simulated turbine is consisted of spiral case, 12 stay vanes, 28 guide vanes and a runner with 4 blades and draft tube.
CFD simulation results show that, the most of local flow parameters like velocities and flow angles at inlet and outlet
are affected by the operating regimes of turbine. It is observed that, the discharge factor, efficiency and specific energy
obtained from CFD simulations are well agreed with experimental results for any axial turbine.
In this study, analytical calculation and numerical simulations were performed together to design a Francis turbine
runner blade. Input parameters of net head and flow rate were chosen 33.53 m and 1,048 m3/s respectively for design
point. Single blade was designed by using Bovet approach. Then creating blade geometry by using Ansys Bladegen
V16.1, numerical simulation was performed by using Ansys CFX solver software. Single blade efficiency obtained
from CFD analysis was validated with the committed blade efficiency of 0.89.

Nomenclature

b guide vane height (-)

H net head of turbine (m)
 lenght (-)
N rotational speed (rad/s)
n specific rotational speed (-)
 modified pressure (Pa)

specific radius on meridional plane (m)
 sum of body forces (N)
 volumetric flow rate (m3/s)
 specific volumetric flow rate (m3/s)
β angle of blade
 viscosity

2.
Design approach

The design approach of Bovet method is based on fundamental parameters of turbine. Those are net head (H),
flow rate () and rotational speed (N) [3]. It is seen that Bovet approach calculates dimensionless specific rotational
speed by using these main variables given in Eq. (1).

N (V
/ π )
1/ 2

n= (1)
(2 gH )3 / 4
Specific rotational speed, ‘n’ is used to calculate the values of runner channel parameters in the meridional plane,
which are shown in Figure 2.
888 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894

Fig. 2. Parameters of runner channel meridional plane [3].

As seen in Figure 2, Bovet approach is capable to calculate runner channel parameters in dimensionless form as
function of ‘n’. Method uses specific radius of ‘

’ which is the distance between rotational axis of the runner and
intersection of shroud and trailing edge of the blade. Specific radius can be calculated from Eq (2).

1/ 3
⎛ V
1 ⎞
R2e = ⎜⎜ ⎟⎟ (2)
⎝ π Nv
2e ⎠
Where v
2 e represents specific volumetric flow rate. It is taken as 0.270 for this study. To obtain dimensional results,
parameters of meridional channel are multiplied with specific radius of the runner. By this way whole variables of the
runner channel are reduced to functions of net head, flow rate and rotational speed.

Table 1. Equations of meridional plane parameters [3].

b0 = 0.8(2 − n)n 0.16

r0i = ymi = 0.7 +
n + 0.08
li = 3.2 + 3.2(2 − n)n le = 2.4 − 1.9(2 − n)n

The correlation between specific rotational speed and meridional channel parameters are not linear as seen in Table
1. So this kind of relation between ‘n’ and the parameters causes logarithmic variations and naturally this leads to
possibility of designing different types of blades given in Figure 3.
After these necessary calculations are completed, runner channel meridional plane can be created. Then,
intersection points of runner blade leading and trailing edges with hub and shroud were determined, respectively.
Therefore, runner blade meridional profile was placed on runner channel meridional plane. It is known that the most
important design parameter of a runner blade is profile throughout leading to trailing edges. For this purpose, Bovet
separates the runner blade with streamlines which have equal flow rate between each other and this is main
characteristic of these streamlines. Seven streamlines were used which are named from hub to shroud respectively i,
i’, i’’, m, e’’, e’, e where “i” spline represents hub curve while “e” curve representing shroud (Fig. 4). It guides to
design more sensitive geometry while constructing runner in a CAD software. Completing construction of meridional
profile, it is also need a determination for designate the angle of specific points on the blade. Recent studies are showed
that the angle of leading and trailing edges are highly significant when designing high efficient runner. The design
approach calculates radiuses and angles at intersection points of streamlines with trailing and leading edges by using
velocity triangles.
 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894 889

Fig 3. Change in blades shape with according to specific rotational speed ‘n’ [3].

Fig. 4. Streamlines on channels' meridional plan [3].

In this study, a Francis type turbine runner that designed with Bovet approach is analyzed by using CFD simulation.
Based on report given by Bovet [3], 33.53 meters net head, 1.048 m3/s volumetric flow rate is chosen while the
rotational speed of the turbine is taken 765 rpm. The committed efficiency is %89 for mentioned runner. By using
these parameters in Eq. (1), the dimensionless rotational speed is calculated as 0.356. From Eq. (2), the R2e is calculated
as 254 mm. The calculated values of radiuses and angles of the trailing and leading edges on seven streamlines are
given Table 2.

Table 2. Radiuses and Angles of Leading and Trailing Edges on 7 Streamlines [3].

Control R1 R2
β1 β2
Lines (mm) (mm)
i 247.6 52̊ 37′ 110.8 36̊ 32′
i′ 247.8 52̊ 25′ 124.4 32̊ 01′
i′′ 248.2 52̊ 00′ 147.1 26̊ 42′
m 249.1 51̊ 07′ 177.8 21̊ 46′
e′′ 252.9 47̊ 37′ 206.9 18̊ 35′
e′ 260.2 41̊ 56′ 242.2 16̊ 34′
e 271.8 36̊ 06′ 254.0 15̊ 03′
890 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894

3.
Numerical study

Numerical simulations were performed to calculate single blade efficiency. In addition, pressure distributions,
streamlines over the blade were figured out. First of all, blade geometry was created by using Ansys Bladegen V16.1.
During this process it is possible to get runner blade channel meridional plane which is given in Figure 5. It is seen
that single blade which is placed on runner channel meridional plane by using Bladegen module has similar profile
with the specified in red frame. Since the specific rotational speed was calculated 0.356 for this study, this similarity
shows that meridional profile of the blade created with Bladegen module is nearly same profile for the value of
dimensionless specific speed 0.4. Three dimensional geometry created precisely, for CFD analysis. 3-D view of the
runner that designed in this study shown in Figure 6.

Fig. 5. BladeGen Meridional Profile.

Fig. 6. 3D Geometry of the Runner.

After completing construction of the runner, grid model has been generated by using Ansys Meshing, with using
fine mesh quality, tetrahedron type mesh. Grid system was consisted of 698.576 elements which were generated with
skewness values below 0.80 and average element quality 0.81 was showed in Figure 7.
 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894 891

Fig. 7. Grid system of one blade flow volume.

In the literature, k-ε and k-ω turbulence models are widely used. There are many CFD based studies which are
validated with the experimental model tests. It is observed that validated numerical studies use k-ε and k-ω turbulence
models to simulate fluid flow in the entire turbine [11]. In the numerical setup, domain material was selected water
was considered to be single phase and incompressible. In this study, k- ε turbulence model have been selected for
steady state analysis. The governing equations given by Ansys Inc, CFX solver theory guide [12] are used. The
continuity equation is;

∂ρ ∂
+
∂t ∂x j
(
ρU j = 0 ) (3)

and momentum equation is;

∂ (ρU i ) ∂ ∂p′ ∂ ⎡ ⎛⎜ ∂U i ∂U j ⎞⎤
∂t
+
∂x j
(
ρU jU i = − +) ⎢ μ eff
∂xi ∂x j ⎢ ⎜⎝ ∂x j
+
∂xi
⎟⎥ + S M
⎟⎥
(4)
⎣ ⎠⎦

where  is the sum of body forces,  is the modified pressure, 

 is the effective viscosity so that:

μ eff = μ + μt (5)

where  is the turbulence viscosity.

The values of k and ε come directly from the differential transport equations for the turbulence kinetic energy and
turbulence dissipation rate:

∂(ρk ) ∂ ∂ ⎡⎛ μt ⎞ ∂k ⎤
∂t
+
∂x j
(
ρU j k =
∂x j
) ⎢⎜⎜ μ +
σk
⎟⎟ ⎥ + Pk − ρε + Pkb (6)
⎢⎣⎝ ⎠ ∂x j ⎥⎦

∂(ρε ) ∂ ⎡⎛ μt ⎞ ∂ε ⎤ ε
+ (ρU j ε ) = ∂ ⎢⎜⎜ μ + ⎟⎟ ⎥ + (Cε 1 Pε − Cε 2 ρε + Cε 3 Pεb ) (7)
∂t ∂x j ∂x j ⎢⎣⎝ σε ⎠ ∂x j ⎥⎦ k

Cε 1 , Cε 2 , σ k and σ ε are constants.

CFD analyses was performed for a single runner blade. The boundary conditions of a single blade of the runner is
described in Fig 8. Guide vanes and stay vanes were not included in the analysis. Inlet flow condition was set as
892 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894

circumferential direction to blade with angle of 12o at radial component while axial component is taken as zero degree.
Outlet was set as opening pressure and velocities are chosen parallel to rotation axis of the turbine. For side walls of
the flow volume were arranged as rotational periodic. Hub and shroud curves and blade surfaces were assumed as
non-slip wall boundary conditions given Table 3.
In CFD applications different types of validations are required to ensure the accuracy of analyze. For this reason
in this study; RMS residuals, flow rate check, net head re-calculation using computed inlet pressure and minimum
absolute pressure controls were used. RMS residuals were set to 10-6 grade and these criteria have been achieved.

Fig. 8. Boundary conditions in 3D view a) side view b) top view

Table 3. Boundary Conditions.

Boundary Area Boundary Conditions

Inlet Mass Flow
Outlet Opening Pressure
Blade Wall
Hub and Shroud Wall
Front and Back Side Rotational Periodic

4. Results and discussion

So far the problem definition and numerical setup of the CFD analysis have been carried out, so in this title the
results of the CFD analysis are submitted. The result of runner analysis is given in Table 4, where residual target is
set to 10-6 degree for converge criteria. Hydraulic blade efficiency is calculated as 0.88 while in example committed
efficiency was given as 0.89. To conclude that, for the very first preliminary approach, the Bovet method guides with
appropriate design procedure.

Table 4. Results of CFD Simulation.

Parameters Calculated Results

Residual Target 10-6
Hydraulic Blade Efficiency 0.88
Torque on 1 Blade 288.92 Nm
Mass Flow Average of Pressure on Inlet 221333 Pa
Mass Flow Average of Velocity on Outlet 13 m/s

Beside the numerical values of the runner, the post process figures should be construed such as total pressure,
absolute pressure, streamlines and velocities at boundaries. For providing high efficiency and balance, pressure in the
turbine blade should be reduced, gradually. In Fig. 9 and Fig. 10 pressure distributions on meridional schemes and
pressure distributions on blade both pressure and suction side are shown, respectively. It can be seen from the figures,
pressure distribution behavior is in an anticipated form. Thus, pressure should decrease from inlet to outlet, gradually.
 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894 893

As seen in meridional plane, on the blade pressure side, pressure variation follows the expected trend. However, the
suction side of the blade, an unexpected pressure drop region have been seen on intersection of trailing edge and
shroud curve. There should be a negative pressure area on the suction side. In this case, the main problem is the
occurrence of negative pressure at the inlet part of suction side, it can cause to cavitation and oscillation (in Fig.10).

Fig. 9. Inlet to Outlet meridional profile a) without blade b) blade presented

Fig. 10. Pressure distribution on the blade a) Pressure Side b) Suction Side

In order to obtain high efficiency, fluid should follow a path without any separation on blade to transfer its potential
energy to runner, effectively. Fluid should enter high pressure and low velocity values at inlet relatively to the outlet.
The streamlines on the runner are given in Fig 11. When the figure is examined, it is clearly seen that the fluid follows
a path which has separation from blade at inlet and near to hub curve. Moreover, it is seen that, the velocity at the
blade leading edge is slower than the trailing edge. It is also observed that velocity distribution from inlet to outlet has
a smooth transition. This is consequence of layer by layer pressure drop.

Fig. 11. Stream lines in 3D view

5.
Conclusion

In this paper, meridional plane of Francis turbine runner is designed based on the Bovet method. Bovet method is
an unsophisticated method that uses empirical relations to determine the dimensions of the blade in the meridional
plane. After determination of the flow streamlines between the entrance and exit edges, other turbine specifications
such as entrance and exit flow angles are calculated. By using of conformal mapping method, the perpendicular view
894 Eyup Kocak et al. / Energy Procedia 111 (2017) 885 – 894

of runner is obtained. After obtaining CAD design of the runner, CFD simulations are employed to compute efficiency
of the runner, pressure distributions on blades and flow characteristics on the blade surface. Turbine runner efficiency
was computed numerically only 1% difference with respect to committed efficiency designed by Bovet method. The
results of the analysis showed that, Bovet design approach can be used to pre-design for an efficient runner. After
preliminary design of the runner obtained by Bovet approach, according to CFD simulation results, modification on
the runner design can be done, so that efficiency and performance of the runner can be increased. In this study it is
shown that, Bovet method is a reliable method for pre-design of Francis turbine runner. Turbine runner is the most
important component effecting turbine efficiency. Nowadays, Francis turbine overall efficiency reaches about 94.5
percent for hydroelectric power plants. CFD simulations become practical and novel method to design high efficiency
runner. Turbine manufactures have to validate prototype turbine guarantied performance values by model turbine test
reports. Although model turbine test validation process stay in importance, reliable CFD simulation results can be
used as a tool to validate prototype turbine performance characteristics, directly. The results of the analysis showed
that, Bovet design approach is able to pre-design for an efficient runner. For the preliminary design by using Bovet
approach, it can be reached about %90 runner efficiency. In designing a proper runner, cavitation formation must be
considered. Designing preliminary turbine runner by Bovet approach, very small cavitation regions are observed on
the turbine runner. With small modifications on the design of runner blade, by considering CFD simulation results,
the cavitation regions can be eliminated, easily. In the literature, empirical relations and graphical results obtained
from experimental data are employed to preliminary design of the turbine runner. Theories developed by Pfleiderer
and Bovet, provide an opportunity to design a runner blade from leading edge through trailing edge. In Bovet
approach, curve of streamlines are determined leading and trailing edges by assuming equal flow rate for each of flow
channels. In this way, a runner blade profile, which has no flow separation, may be designed with all the angles and
thicknesses up from leading to trailing edges. CFD is a cost effective computational tool for flow simulation and
investigation for hydraulic turbines and can provide detailed flow information. This information can be used to modify
the preliminary design obtained by Bovet approach, so that efficiency and performance of the runner can be increased.
To challenge and improve Bovet design method in further studies, cavitation controls and wall-shear post processes
should be investigated. In addition to that, blade thickness, angles and profiles can be optimized. Beside these studies,
whole turbine analyses that included spiral casing, stay vanes, guide vanes and draft tube, might be considered to
study of pressure behaviors and also their effects on efficiency. Ultimately, guide vane position optimizations, draft
tube pressure gain optimizations, vortex analyses, spiral casting pressure accumulations might be studied.

References

[1] U.S. Geological Survey Statistics database [Online]. Available: https://water.usgs.gov/edu/wuhy.html.

[2] Penche C, Minas I. Layman's handbook on how to develop a small hydro site. 2nd ed. European Commission:
European small hydropower association; 1998. p. 75-76.
[3] Bovet T. Contribution to the study of Francis – Turbine Runner Design. Lausanne; 1963.
[4] Miloş T, Bârglâzan M. CAD technique used to optimize the Francis runner design. The 6th International
Conference on Hydroulic Machinery and Hydrodynamics 2004.
[5] Choi H, Zullah MA, Roh H, Ha P, Oh S. CFD validation of performance improvement of a 500 kW Francis turbine.
Renewable Energy 2013;54:111-123.
[6] Zhang H, Zhang L. Numerical simulation of cavitating turbulent flow in a high head Francis turbine at part load
operation with OpenFOAM. Procedia Engineering 2012;31:156-165.
[7] Anup KC, Thapa B, Lee Y. Transient numerical analysis of rotor-stator interaction in a Francis turbine. Renewable
Energy 2014;65:227-235.
[8] Thapa B, Thapa B, Eltvik M, Gjosater K, Dahlhaug O. Optimizing runner blade profile of Francis turbine to
minimize sediment erosion. The 26th IAHR Symposium on Hydraulic Machinery and Systems. 2012.
[9] Obrovsky J, Krausova H, Spidla J, Zouhar J. Development of high specific speed Francis turbine for low head
HPP. Engineering Mechanics 2013;20:2:139-148.
[10] Prasad V. Numerical simulation for flow characteristics of axial flow hydraulic turbine runner. Energy Procedia
2012;14:2060-2065.
[11] Zeng Y, Liu X, Wang H. Prediction and experimental verification of vortex flow in draft tube of Francis turbine
based on CFD. Procedia Engineering 2012;31:196-205.
[12] Ansys Inc. CFX solver theory guide. Release 2013;15:79-90.